How can I match namespace prefixes with regular expression

Question

What I really want to do is to match all namespace prefixes both for tags and attributes by using a regular expression.

For example:

<foo:person bar:id="43">
  text text <nomatch:free text
  <baz:name>
    text nomatch:free> text
  </baz:name>
    text nomatch:free="44" text
    <age bae:years="44" />
    text
</foo:person>

What I want is to extract only the following prefixes: foo, baz, bad, bae, bar

I cannot find a regular expression for implementing this.

Please take account that the regular expression should not have hardcoded the names foo, baz, bad, bae, bar...

Answer 1

Something like this?

[<]([^:]+)

will put the prefix (after the angle bracket and before the colon) into group 1.

I see that I misinterpreted the question, after looking more closely.

[<](((\w+):\w+)(\s+(\w+):\w+="\d+"\s*[\/])?)>

which will have a result in groups $3 and possibly $5. I think this will do it.

Whoops, missed a ?

 [<](((\w+):\w+)(\s+(\w+):\w+="\d+"\s*[\/]?)?)>